In man, normal urinary bladder contractions are mediated, (inter alia), through cholinergic muscarinic receptor stimulation. Muscarinic receptors not only mediate normal bladder contractions, but may also mediate the main part of the contractions in the overactive bladder resulting in symptoms such as urinary frequency, urgency and urge urinary incontinence.
After administration of fesoterodine and other phenolic monoesters of formula (III) to mammals, such as humans, these compounds are cleaved to form the active metabolite. The active metabolite is known to be a potent and competitive muscarinic receptor antagonist (WO 94/11337). Therefore, fesoterodine and other phenolic esters of formula (III) represent potential prodrugs for the active metabolite, and are drugs which are effective in the treatment of overactive bladder with symptoms of urge urinary incontinence, urgency, and urinary frequency, as well as detrusor hyperactivity (as described e.g. in U.S. Pat. No. 6,713,464). Tolterodine is another well known drug for the treatment of overactive bladder.
Two different routes for the synthesis of the phenolic monoesters of formula (III) such as fesoterodine have previously been described in U.S. Pat. No. 6,713,464 and WO 01/96279, respectively. WO 01/49649 discloses a certain method for producing tolterodine.
The synthesis of the active metabolite is also known in the prior art. WO 94/11337 and WO 98/43942 both describe a multi-stage process to synthesize the active metabolite.
However, all these prior art processes are inconvenient, because they comprise a large number of steps e.g. in accordance with the synthesis disclosed in WO 94/11337, 11 steps are necessary for obtaining the active metabolite. Similarly, 12 different reaction steps are necessary for producing the phenolic monoesters of formula (III) (see U.S. Pat. No. 6,713,464).
A first approach for shortening the synthesis of the phenolic monoesters of formula (III) is disclosed in WO 01/96279. In the process according to WO 01/96279, the preferred R-enantiomer of the compounds of formula (II) or (III) is obtained by utilizing the diastereomeric cinchonidine salt of (R,S)-4-phenyl-2-chromanone-6-carboxylic acid ((2b), Scheme 1). If this salt is crystallized, the R-enantiomer of 4-phenyl-2-chromanone-6-carboxylic acid predominates as the acid component (more than 95% ee.). By recrystallization, the enantiomeric purity can be increased up to 99% ee.
The optically pure lactone (step 3, (3)) is then liberated by acidification and subsequently converted into its methyl ester (4). The lactone (4) is then reduced with one molar equivalent of a hydride, thereby obtaining the lactol (5). The lactol intermediate (5) is then used to prepare the active metabolite (II) in 2 additional steps by first reductively aminating the lactol (5) and in a second step by the reduction of the ester substituent to give the benzylic hydroxyl function of the active metabolite (II), which can be then acylated to give a compound of formula (III).
Notwithstanding the significant reduction in the number of necessary operations as compared to previous routes, the synthesis of the active metabolite of formula (II) still requires 8 steps in total, and 5 steps from 4-phenyl-2-chromanone-6-carboxylic acid.

Owing to the large number of steps involved, all prior art processes are complex, and the overall yield of the active metabolite is unsatisfactory. As a consequence there was a need for a further shortening of the synthesis of the compounds of formula (II) or (III), whereby the above disadvantages may be avoided.
In WO 01/96279 the reduction of the lactone (4) was performed at mild conditions and with a stoichiometry [reducing agent/compound of formula (4)] of about 1:1 or less because of concerns that harsher reduction conditions would lead to an opening of the lactone ring. Based on the state of the art, it would have been expected that harsher reduction conditions would result in an over-reduction of the intermediate lactol such that both the lactone as well as the benzoate ester functions would be fully reduced thus leading to the synthetically unwanted primary alcohol depicted below (March's Organic Chemistry, 5th Ed, Wiley Publication, 2001, see particularly tables 19-3 and 19-5; Walker, Chem Soc Rev 5, 1976, 23; see particularly Table 7; Soai et al, J Org Chem 51, 1986, 4000).
Expected Reduction Reaction:

The problem was thus to achieve the reduction of the benzoate ester or benzoic acid function to the primary alcohol while at the same timing stopping the reduction of the lactonic ester at the aldehyde (lactol) stage, with the particular challenge that lactones are generally more susceptible to reduction than carboxylic acids or esters (see e.g. March's Organic Chemistry, 5th Ed, Wiley Publication, 2001, tables 19-3 and 19-5). Surprisingly, it has now been found that under appropriate conditions in fact the benzoate ester of compound (4) can be selectively reduced without reduction of the lactol. The resulting compound of formula (I) can then be converted in one step to the active metabolite of formula (II), thereby saving one step in the overall production process (scheme 2).
